Field of the Invention
The present invention generally relates to a vehicle localization system for an autonomous vehicle. More specifically, the present invention relates to a vehicle localization system for an autonomous vehicle that selectively switches between two localization systems based on predetermined conditions.
Background Information
As understood in the art, autonomous vehicles generally include a vehicle control system that is capable of performing driving operations to guide the vehicle to a predetermined destination without input from the driver or with only minimal input from the driver. The vehicle control system typically uses a plurality of high speed cameras to monitor areas surrounding the vehicle. The cameras detect, for example, traffic lights, signs, intersection markers, other vehicles and obstacles, and provide visual data to the vehicle control system.
Conventional vehicle localization systems compute the position of a vehicle by comparing an image of a stored three dimensional map and a camera image. In particular, an edge image is extracted from an actual image acquired by a vehicle camera with which the vehicle is equipped. The position and attitude angle of the vehicle camera is adjusted so that a virtual image from a three dimensional map which recorded the position and type of edge of the environment by three dimensions is projected on the positional attitude of the vehicle camera. Accordingly, the position and attitude angle in three dimensional space of the vehicle camera can be estimated.
Moreover, successive images from cameras can be compared to determine the movement of the vehicle. Specifically, by comparing the location of a plurality of matching pixels from successive images, distance information can be obtained. The distance information can be compiled to determine movement of the vehicle in various directions and angles.